Image Downsampling
Also known as: downsample image, resize for storage, image downsampling
Image downsampling reduces a photo's pixel dimensions, computing fewer output pixels from many input pixels. Because file size scales with pixel count, halving width and height cuts the pixel total to roughly a quarter, which is how compressors shrink photos for storage.
- File size scales with pixel count, so halving width and height cuts pixels to about one quarter.
- Quality depends on the resampling filter, e.g. bilinear, bicubic, or Lanczos.
- Distinct from compression: downsampling changes resolution, compression discards detail at the same size.
How downsampling works
Downsampling maps a high-resolution grid of pixels onto a smaller one. Each output pixel is computed from a neighborhood of input pixels using a resampling filter such as bilinear, bicubic, or Lanczos, which average and weight nearby samples to avoid jagged edges and aliasing. Good downsamplers apply a low-pass step first so fine detail is blended rather than dropped abruptly.
Pixel count grows with the product of width and height, so the savings are nonlinear: scaling a 4000x3000 photo down to 2000x1500 keeps half the dimensions but only about a quarter of the pixels. Fewer pixels means less data for the encoder to store, which is why downsampling is one of the biggest levers for shrinking an image.
Downsampling vs. compression
Downsampling is distinct from image compression. Compression (for example JPEG quality settings) keeps the same dimensions but discards information the eye is less sensitive to. Downsampling changes the resolution itself. Most tools combine both: shrink the dimensions to a sensible cap, then re-encode at a chosen quality.
For phone storage this matters because original camera photos are far larger than any screen can display. Resizing a multi-thousand-pixel image down to something closer to a display or sharing size, then re-encoding to HEIC, JPEG, or WebP, can reclaim substantial space while still looking sharp on a phone. That is the core mechanism behind features that optimize photos and the resizers a cleaner pairs with.